SyNET SIGNED

Project "SyNET" data sheet

The following table provides information about the project.

 Partnership

Take a look of project's partnership.

Map

 Project objective

Nearly half of Japan without Internet for hours; thousands of airline passengers grounded; 600 emergency calls not received. These are just three recent illustrations of the consequences of human-induced network misconfigurations. Such errors are caused by a fundamental gap: distilling the high-level requirements network operators have in mind into complex, low-level device configurations. In 2018, network operators are still trying to bridge this gap manually.

The goal of this proposal is to address this gap by automatically synthesizing low-level configurations out of high-level requirements. This is a hard, inter-disciplinary challenge requiring development of new fundamental methods spanning multiple scientific fields including computer networks, programming languages, and probabilistic reasoning. Our vision is akin to program synthesis, a new programming paradigm in which “one tells the computer what to do, not how to do it”.

We propose to address this challenge via three work packages: (i) user-aware synthesis where we develop techniques to facilitate how operators communicate their intents and define the space of configurations to search over; (ii) robustness-aware synthesis where we develop synthesis techniques that can enforce probabilistic properties (e.g., minimum reliability level) in uncertain network environments; and (iii) performance-aware synthesis where we develop methods to synthesize more performant configurations (e.g., to make the network converge faster).

We believe SyNET will trigger a paradigm shift and spearhead the next generation of network management tools blending formal methods, probabilistic models, and user-intuitive interfaces. By alleviating the need to configure networks, SyNET promises to avoid the vast majority of downtimes plaguing today’s networks and to make tomorrow’s networks more flexible and, ultimately, more efficient.